IETF
iccrg
iccrg@jabber.ietf.org
Tuesday, November 15, 2016< ^ >
jishac has set the subject to: Internet Congestion Control Research Group
Room Configuration
Room Occupants

GMT+0
[06:37:58] Meetecho joins the room
[06:45:11] Marcus SUN joins the room
[06:45:12] John Border joins the room
[06:45:12] hiren panchasara joins the room
[06:45:12] Randell Jesup joins the room
[06:45:28] Benjamin Hesmans joins the room
[06:46:36] Karan Verma joins the room
[06:46:58] Samuel Jero joins the room
[06:47:17] Samuel Jero_ joins the room
[06:47:33] Praveen Balasubramanian joins the room
[06:51:11] Ingemar Johansson joins the room
[06:55:04] Ingemar Johansson leaves the room
[06:55:09] Ingemar Johansson joins the room
[07:05:23] Karan Verma leaves the room
[07:05:28] Karan Verma joins the room
[07:10:33] Simon Romano joins the room
[07:11:30] Praveen Balasubramanian leaves the room
[07:11:42] Praveen Balasubramanian joins the room
[07:14:15] m h joins the room
[07:21:45] Roland Bless joins the room
[07:26:02] Roland Bless leaves the room
[07:31:19] Jana Iyengar joins the room
[07:33:40] <Jana Iyengar> Is this alive?
[07:33:56] jholland@lightwitch.org joins the room
[07:34:17] <Jana Iyengar> Can someone else ping here just so we can confirm that this room is working?
[07:34:25] <Ingemar Johansson> ping
[07:34:30] <Randell Jesup> pong
[07:34:35] Tim Chown joins the room
[07:34:39] <Jana Iyengar> Thanks Ingemar and Randell!
[07:35:00] <Meetecho> note to chairs: Randell Jesupin virtual queue
[07:35:02] <Jana Iyengar> Randell, do you have a question?
[07:35:16] Mirja Kühlewind joins the room
[07:35:27] <Tim Chown> he's in the queue...
[07:35:34] <Jana Iyengar> yup
[07:36:06] Mirja Kühlewind leaves the room
[07:36:12] Mirja Kühlewind joins the room
[07:36:38] <Jana Iyengar> Yeah, sorry, just confirming that he just showed up, because neither of the chairs noticed earlier. The meetecho comment clarifies this, so all good.
[07:44:54] Karan Verma leaves the room
[07:44:56] Karan Verma joins the room
[07:50:34] jholland@lightwitch.org leaves the room
[08:00:50] Roland Bless joins the room
[08:05:52] Padma Bhooma joins the room
[08:05:58] <Ingemar Johansson> Do the flows probe the minRTT at the same time ? (31s, 41s)
[08:06:02] <Praveen Balasubramanian> How do you deal with RTT probing for short-lived flows (say that last less than 10 seconds)? What avoids synchronized probe RTT modes across different connections?
[08:09:54] <Praveen Balasubramanian> Is RACK+TLP used in conjunction with BBR to deal with packet loss? Is RTT and B/W info discarded after timeout?
[08:09:56] bless joins the room
[08:15:32] Mario Hock joins the room
[08:23:14] zhenjie deng joins the room
[08:24:56] <Mario Hock> Does BBR has an estmate of the capacity of the path, or of it's fair share on the path?
[08:31:48] <Roland Bless> Much better approach to deal with congestion, but what concerns me is that it doesn't react to packet loss. So in severe congestion situations it doesn't back off? What about prevention of congestion collapse?
[08:33:00] Simon Pietro Romano joins the room
[08:37:08] <Meetecho> note to chairs, in case they're not aware of them: Randell Jesup has been in virtual queue for a while
[08:37:18] <Meetecho> *them=this
[08:37:25] <Jana Iyengar> yup getting there. sorry missed it earluer
[08:43:52] <Randell Jesup> meetecho: Upstream delay (me to the room) seems to be REALLY long; on the order of many seconds.....  I haven't seen this in previous IETFs
[08:44:28] <Meetecho> Randell Jesup: we've noticed, not sure what may cause it; the only thing I've noticed is that you're on TURN
[08:44:30] <Randell Jesup> Also there seems to be a few-second delay for audio and video downstream (though not or less for the slides)
[08:45:02] <Meetecho> can you try to do a "reconnect audio"? there's a button for that below, next to the recording label; that may sync it again, if not for just a while
[08:45:18] <Randell Jesup> TURN *shouldn'* add more than a bit of RTT...
[08:45:20] Polina Goltsman joins the room
[08:45:20] <Randell Jesup> ok
[08:46:29] <Randell Jesup> From the gap and conversation when I did that, I don't think it changed much
[08:48:02] <Randell Jesup> If you're on a non-symmetric NAT or open IP at the server end, a TURN server "should" never be needed.  Though that doesn't mean a TURN candidate might not "win" in some cases
[08:49:19] Karan Verma leaves the room
[08:50:46] Karan Verma joins the room
[08:50:54] <Meetecho> just checked again and you're not on TURN, right: not sure why I thought that, I might have confused you with another user
[08:51:01] <Randell Jesup> meetecho: all 3 peerconnections show no TURN in use
[08:51:16] <Meetecho> apart from that, in our server internals I only see a lot of nacks for the speaker video feed
[08:51:43] <Meetecho> and no queued packets on any of your streams, which means we're shooting them right away
[08:52:47] <Meetecho> you can try restarting the speaker video (hovering over it and clicking the refresh icon) to see if that makes it sync again
[08:53:05] <Meetecho> not sure on why you're loosing so many packets from that stream though, as it's not a high bitrate feed
[08:53:42] <Randell Jesup> I don't seem to be seeing any obvious loss
[08:54:57] <Randell Jesup> Biggest problem *seemed* to be upstream, but it can be hard to separate where the delay is.... I presume my downstram is delayed because I see slides change several seconds *before* the user asks for a slide change
[08:55:34] <Randell Jesup> A/V are pretty synced, so that's not it
[08:56:01] <Meetecho> synced as in equally delayed?
[08:56:21] <Meetecho> (especially if slides come so much sooner)
[08:57:19] <Randell Jesup> yes.  a/v are well-synced
[08:59:26] <Randell Jesup> slides change maybe 10-20 *seconds* before the person says "Next"
[09:00:17] <Randell Jesup> I'm going to leave and rejoin to see what happens.
[09:00:58] <Randell Jesup> I'm back...
[09:01:29] <Randell Jesup> Since I'm using the slides as a timing mechanism, and he's on Questions, I can't easily tell the delay now
[09:01:44] <Meetecho> anything weird you could spot in the browser internals about those PCs?
[09:01:59] <Randell Jesup> If you have a low-latency feed, you could send a jabber message when someone says something in  particular
[09:02:37] <Randell Jesup> no.  Low bitrate, no apparent loss when I looked (180Kbps, 0%)
[09:03:58] <Randell Jesup> i think delay is lower now...  I'll watch for any more calls for slide changes.   Also, he was moving his mouse ptr over something and saying it, and there was no apparent delay!.
[09:04:33] <Randell Jesup> rejoining may have fixed it (not too surprising); question is whether delay would build up again
[09:04:35] <Meetecho> I still see many NACKs for the speaker video, though, which means retransmissions may be occurring there
[09:04:54] <Randell Jesup> nacks from me, or in general
[09:05:09] <Meetecho> nacks from you, as that's a recvonly stream for you
[09:05:22] <Meetecho> so we're retransmissing those packets
[09:05:44] <Meetecho> no NACKs for the audio and slides feeds instead
[09:06:59] <Meetecho> not sure if you have a way to verify this Firefox's internals
[09:07:04] <Meetecho> *in
[09:07:12] <Randell Jesup> Slide flow has a weird alternation between old/new/old/new for a while on every slide change
[09:07:41] <Meetecho> that's likely a hiccup in the capture card here, so "normal" :)
[09:10:17] <Randell Jesup> ok.  I'll watch for this in the future.  No delay right now
[09:10:26] <Meetecho> ack!
[09:10:30] <Randell Jesup> switching to netvc for a bit
[09:14:03] Eneko Atxutegi joins the room
[09:14:39] <Meetecho> I think netvc ended already
[09:20:09] albert tommasen joins the room
[09:22:12] Samuel Jero_ leaves the room
[09:22:33] Samuel Jero leaves the room
[09:23:41] Padma Bhooma leaves the room
[09:23:45] <Karan Verma> display problem
[09:24:33] Simon Pietro Romano leaves the room
[09:25:21] Simon Romano leaves the room
[09:25:44] Eneko Atxutegi leaves the room
[09:25:44] Mirja Kühlewind leaves the room
[09:25:44] Ingemar Johansson leaves the room
[09:25:44] hiren panchasara leaves the room
[09:25:44] Randell Jesup leaves the room
[09:25:44] albert tommasen leaves the room
[09:25:44] Marcus SUN leaves the room
[09:25:44] Tim Chown leaves the room
[09:25:44] Mario Hock leaves the room
[09:25:44] Karan Verma leaves the room
[09:25:44] John Border leaves the room
[09:25:44] Praveen Balasubramanian leaves the room
[09:25:44] Roland Bless leaves the room
[09:25:44] zhenjie deng leaves the room
[09:25:44] Polina Goltsman leaves the room
[09:25:44] m h leaves the room
[09:29:54] Meetecho leaves the room
[09:30:31] Benjamin Hesmans leaves the room
[09:39:40] Jana Iyengar leaves the room
[09:43:25] Jana Iyengar joins the room
[09:51:34] bhesmans joins the room
[10:13:50] Jana Iyengar leaves the room
[10:22:45] bhesmans leaves the room
[10:47:13] jholland@lightwitch.org joins the room
[10:56:46] jholland@lightwitch.org leaves the room
[11:29:30] bhesmans joins the room
[13:29:37] jholland@lightwitch.org joins the room
[15:15:13] bhesmans leaves the room
[22:31:58] jholland@lightwitch.org leaves the room
[22:54:59] jholland@lightwitch.org joins the room
[23:04:11] jholland@lightwitch.org leaves the room
[23:57:42] jholland@lightwitch.org joins the room
Powered by ejabberd - robust, scalable and extensible XMPP server Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!